Loss of instrumental activities such as telephone use and money management are inevitable consequences of Alzheimer's disease (AD). In prior work, we have investigated loss of treatment consent capacity in AD using psychometric and physician judgement assessment. However, other important functional capacities, in particular financial capacity, remain unstudied. Little is known about the natural history of loss of financial capacity in AD, and the neurocognitive changes mediating this loss. We propose to investigate these questions longitudinally by recruiting over study years 1-4 a baseline sample of 100 older controls and 100 mild AD patients, and following these subjects annually for the remainder of the five year study (one to four year follow-up). Financial capacity will be conceptualized as a multi-dimensional construct represented by six domains of knowledge: (1) Basic Monetary Skills, (2) Conceptual Knowledge, (3) Cash Transactions, (4) Checkbook Management, (5) Bank Statement Management, and (6) Financial Judgement. Each domain is operationalized by specific financial tasks which are conceptualized along two dimensions: core knowledge type (declarative procedural, judgement) and processing level (automatic, controlled). In Aim 1, we will use this conceptual model and two assessment methods (psychometric and informant report) to investigate loss of financial knowledge to AD. We hypothesize that mild AD patients at baseline assessment will perform equivalently with controls on automatic tasks, but will demonstrate impaired performance on controlled tasks. As dementia progresses, AD patients will demonstrate more rapid decline on controlled versus automatic tasks, and will demonstrated impaired performance relative to controls on automatic tasks. In Aim 2, we will investigate neurocognitive changes associated with loss of financial knowledge in AD. We hypothesize that changes in semantic memory and executive function will predict declining AD patient performance on controlled tasks, and that changes in attention will predict decline on automatic tasks. This study will reveal the structural breakdown over time of financial abilities in AD, and neurocognitive markers for these clinical changes.